Cross-Tabs and Combined Survey Analysis in Reporting¶
"68% of respondents are satisfied" is a headline. It's rarely the finding you actually act on. The more useful version of that number is almost always a comparison - satisfaction among new customers versus returning ones, or across plan tiers, or by region. That comparison is what a cross-tab gives you.
What a Cross-Tab Widget Does¶
A cross-tab breaks one question's results down by the answers to a second question. Instead of a single bar per answer option, you get a grid or grouped chart - each row or series representing one segment, so you can see at a glance whether satisfaction, say, actually differs by plan tier or just looks like it does in the overall number.
Any two questions in the same survey can be cross-tabbed against each other, and a cross-tab widget updates live as new responses come in, the same as a single-question widget does.
Choosing What to Cross-Tab¶
The question you're analyzing goes on one axis; the question you're segmenting by goes on the other. A good segmenting question is usually something respondent-level and relatively stable - plan tier, role, region, how they answered a screening question - rather than another opinion question that's likely to move for the same underlying reasons as the one you're analyzing. Cross-tabbing two opinion questions against each other can still be useful, but read it knowing you're looking at a correlation between two answers, not a segment breakdown.
Combined Survey: Cross-Tabs Across More Than One Survey¶
A single survey's cross-tab compares segments within one dataset. Combined Survey extends the same idea across multiple surveys - useful for a tracking study run quarter over quarter, or for comparing results from two audiences you surveyed separately rather than in one combined instrument.
Once surveys are linked in a Combined Survey view, you can build the same kind of comparison widgets you would within a single survey, with each source survey treated as its own segment. This is the tool to reach for when you want "how did this metric move between Q1 and Q2" or "how do these two audiences differ," rather than manually exporting two surveys' results and reconciling them yourself.
When Not to Cross-Tab¶
A cross-tab needs enough respondents in each segment to mean something - a comparison built on a handful of responses per segment will look dramatic and tell you very little. If a segment is thin, it's usually more honest to report the overall number with a note on sample size than to present a cross-tab that implies more precision than the data supports.
FAQ¶
Can I cross-tab a question against itself?
No - a cross-tab needs two distinct questions, one to analyze and one to segment by.
Does a cross-tab work with open-ended questions?
Cross-tabs are built for questions with defined answer options (single-select, multi-select, scale, and similar). An open-ended question needs to go through Text Analytics first to get categorized answers before it can be used as either side of a cross-tab.
How many surveys can I combine in Combined Survey?
There's no fixed cap on the number of surveys you can bring into a Combined Survey view - the practical limit is how meaningful the comparison stays as you add more sources, not a hard system limit.
Do the surveys in a Combined Survey view need identical questions?
They don't need to be identical, but a comparison is only meaningful where the surveys share a genuinely comparable question - Combined Survey doesn't attempt to reconcile differently worded questions into the same comparison automatically.
For more on picking the right chart type once you know what you're comparing, see Choosing the Right Widget for Your Report.